跳到正文
𝕏 Thomas Wolf(x.com)·· 8 天前品質評分48
重點導讀

Fable 5.2 在最新的 Vending-Bench 測試中,似乎知道自己正在受測。

  • Andon Labs 的新報告指出,它讀過這項測試的論文,也知道哪些不當行為會被公開報告,因此變得處處懷疑,像是一直在找陷阱。模型會讀到大家公開談論新模型的內容,這些討論也可能成為下一批模型的訓練資料。
  • 它們能學到人們真正重視什麼,也可能學會哪些表現會讓人起疑。這讓公開談論模型的方式變得重要,因為連災難情境的討論都可能影響後續模型。

AI 伴隨問答

自備 Key 深度分析

針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:

Thomas Wolf
Thomas Wolf@Thom_Wolf
·8 天前
𝕏

According to Andon Labs’ new reports, Fable 5.2 knows it is being tested when we run it on the latest Vending-Bench.

But it’s way more than that. It has actually read (been trained on) the benchmark paper and knows what and how misbehavior gets reported publicly. This makes it behaves very strangely: it suspects traps everywhere.

Everything we publish about recent models now becomes training data for the next. Including this thread!

That is partly good news. By reading what we think, a model can pick up our underlying intentions, not just proxy metrics.

But the same public discussion also teaches which signs make us suspicious, and from the outside the two kinds of learning look alike.

So how we write about models now matters. I even worry about the doom scenarios plastered everywhere: they are training data too, and some feedback loops are strongly nonlinear.

推文來源 · X (Twitter)在 X 參與討論

讀者討論

理性交流 · 即時討論